Health literacy is commonly operationalized by tests of word pronunciation or reading comprehension. The causal connection between these abilities and health outcomes has yet to be elucidated. Research into the relationship between health literacy and outcomes in HIV care has produced conflicting results. We believe that resolving this confusion requires understanding the actual explanatory models of people living with HIV about HIV disease, ARV action, and other constructs relevant to treatment. We know little about how people with HIV choose to initiate or continue a medication regimen, or to take individual doses, along with other health-related behaviors. Until we fill this gap in our knowledge, we cannot design evidence-based educational strategies to give people with HIV the knowledge and understanding to support effective self-care and informed decision making. The long-term goal of this research is to apply new understanding and tools to the development and testing of educational interventions to support improved self-care and outcomes in HIV. The overall objective of this R21 is to understand the explanatory models of persons with HIV about the disease and its treatment, and in particular to identify models that conflict in potentially harmful ways with biomedical understanding. We also propose to develop and pilot test a novel HIV literacy assessment tool based on this broader concept of literacy. Future research will examine the predictive validity of the instrument for ART adherence and elements of physician-patient communication, and use it to develop educational tools to improve HIV-specific health literacy. Our central hypothesis is that individuals whose conceptual models accord more closely with biomedical understanding are more likely to act in accordance with medical advice. The rationale for the approach we propose is that a broadened evidence base is needed to expand the concept of health literacy to become more relevant to the clinical care of persons with chronic diseases. We have two Specific Aims: 1) Develop an in-depth understanding of common patterns of cognition about HIV and HIV treatment among people living with HIV. We will conduct in depth interviews with a diverse sample of 50 persons with HIV regarding their knowledge, beliefs, and conceptual models, and use a grounded theory approach to analyze the interviews to elucidate a taxonomy of cognitive models and heuristics for treatment decision making. 2) Based on the findings from Aim 1, we will design and cognitively test questionnaire items to assess respondents' HIV content literacy, and categories of explanatory models and decision heuristics, and preliminarily assess the relationship of response patterns to medication adherence with a diverse sample of 250 people living with HIV. This new understanding (Aim 1) can have tremendous impact because it uses rigorous qualitative methods to develop a patient-centered evidence base for HIV treatment education; while the assessment instrument (Aim 2) will translate the findings into a practical, clinical tool. Many of the lessons learned will be generalizable to other chronic conditions such as depression, diabetes, and hypertension.